Memory Genome Engine Offers Structured, Local-First Memory Layer for AI Agents

A developer has built Memory Genome Engine, a Rust-based memory subsystem designed to give AI agents durable, structured memory across multiple sessions. Unlike conventional chat history, the system stores each memory item as a structured cell with explicit markers for type, scope, trust level, and sensitivity. The tool is local-first, requiring no hosted cloud service, so users can inspect and control exactly what the agent remembers. It supports command-line interactions that let users store specific facts and later retrieve only the context most relevant to a given task. The project is open source and available on GitHub, aiming to replace opaque memory handling with a transparent, inspectable alternative.